Quantum Computing Has a Noise Problem

Quantum computers have huge problem. Or, to be more precise, a lot of incredibly tiny ones. These futuristic devices promise to revolutionize everything from the financial industry to drug discovery by harnessing the power of quantum uncertainty—instead of using bits like your laptop or phone, quantum computers use qubits, which means they can perform certain tasks. much faster than traditional computers and can better simulate natural processes.

Tech giants including Google, Microsoft, and IBM are rushing to create quantum devices, but overall, the field is mired in an era known in the business as “intermediate noise floor,” or NISQ. Today’s quantum computers are fragile devices that can be thrown off course by the slightest interference from the environment: they are slow, small and not very accurate, which means that right now they are useless.

Sabrina Maniscalco hopes to change that. She is the co-founder and CEO Algorithmic, one of the few startups developing software for noisy quantum computers that we have access to today. “Software and algorithms for near-future devices are key to discovering and discovering useful industrial applications,” she says.

The company grew out of research at the University of Helsinki, where Maniscalco is a professor of quantum information, computing and logic, after working in South Africa, Edinburgh and her native Sicily. “It all started with us trying to find the best applications for these very noisy quantum computers early on,” she says.

They settled on the problem of “noise”. Algorithmiq is developing ways to counteract the noise that quantum computers suffer from: not cooling fan noise, but tiny environmental changes that can take qubits out of a delicate state called superposition. It is this state, which can be roughly represented not as 0 or 1, but as both at the same time, that makes quantum computers so powerful, but at the same time so difficult to create.

Algorithmiq develops sophisticated noise modeling and reduction techniques so that early stage devices can be used for experimentation. For starters, the company has focused on chemical modeling, a promising potential use case for quantum computers as they mimic the uncertainties of nature. He compares his noise reduction algorithms by simulating molecules such as dichromium, which are simple enough to be modeled with today’s quantum computers, but complex enough to demonstrate the capabilities of these devices. Algorithmiq, which recently announced a partnership with IBM, plans to apply the same principles to more complex structures with potential applications, including drug discovery in the pharmaceutical industry, Maniscalco says. “We think of ourselves as the first quantum biotech company,” she says.

This article was originally published in the January/February 2023 issue of WIRED UK magazine.